Search

The problem of fatigue characterization of materials by small data samples is here considered and a Bayesian procedure for data analysis is presented. For both hot rolled and quenched and tempered carbon steels, on the basis of some literature data and by well-established correlations between fatigue and tensile properties, a prior probability density function, able to model most of the available information on material fatigue strength, is defined. Prior information, when fused by means of Bayes theorem with information from experimental data, allows to identify with greater accuracy the fatigue properties of the tested steel A Montecarlo study is carded out to verify the effectiveness of the proposed method, generating small samples of fatigue data by virtual testing. [ABSTRACT FROM AUTHOR]